It's not that more money would make you inherently smarter, but rather that more intelligent parents might (statistically) both make more money and also have more intelligent children.

Those children might go on to score better on the SAT, leaving a correlation between parental income and SAT scores that would be unsurprising at each step.

Your point about inherited wealth is not demonstrated on the chart you showed, which was parental incomes (not wealth) and in ranges unlikely to be perturbed by vast inheritances.

> Also, if you want to measure IQ, why not use an IQ test and not some arbitrary proxy?

I don't want to; I suspect colleges and D. E. Shaw felt like SAT scores were an acceptable/practical proxy for whatever purpose it was that they had. I thought it was weird and slightly off-putting to be asked, but it seems to work for them.